Drug resistant tuberculosis (TB), especially multi-drug resistant TB (MDR-TB), is a major public health threat. The prevalence of MDR-TB is increasing, including recent reports that indicate that some cases of TB are now resistant to all known anti-TB drugs. Treatment of MDR-TB is less effective than for drug susceptible TB, is associated with more severe drug toxicities, is more expensive, and for injectable drugs requires hospitalization. Treatment of MDR-TB in children, in whom TB disease is very different from that in adults, is poorly characterized, if at all. Our group has used the hollow fiber system model of TB (HFS) to show that one of the main reasons for acquired drug resistance (ADR) and therapy failure is the between-patient differences in drug metabolism. In some patients who rapidly and differentially metabolize some drugs, there is created a situation whereby they are effectively under a single drug that is effective, which leads to ADR. Another reason for ADR is that some Mycobacterium tuberculosis (Mtb) lineages have been shown to be hypermutable. In treatment of MDR-TB, a third problem is the high toxicity of drugs, leading to poor completion of therapy. We successfully retrofitted the HFS with 3-dimensional human organotypic liver and skin tissue, so that drug toxicities can be examined at the same time we examine efficacy of regimens against MDR-TB. In addition, we have designed a HFS that is relevant to disseminated intracellular disease in children. Our aim is to design an anti-TB regimen comprising of drugs that are (a) off-patent, (b) cheap, (c) readily available, and (d) can be administered by mouth, in the HFS that will be effective for different Mtb lineages. Efficacy results will be validated using two mouse models, with drug pharmacokinetics similar to those in human adults and children. We will also investigate measuring drug concentrations followed by an iterative intervention with individualized dosing to improve efficacy, reduce ADR, and reduce toxicity, while still being cost-effective in resource poor settings. We will attain our goals by performing a series of HFS experiments for efficacy, ADR, and toxicity in Dr. Gumbo's laboratory at UT Southwestern Medical Center. Next, we will validate the optimal drug regimens in two mouse models, one for adult pulmonary TB and the other for disseminated TB, based on humanized pharmacokinetics in Dr. Eric Nuermberger's laboratory at Johns Hopkins Medical Center. The effects of treatment with chosen drugs on mutation rates of different Mtb lineages will be measured in Dr. Sarah Fortune's lab at Harvard. Results will be employed in computer aided clinical simulations in order to translate dosing strategies to patients.